long-term care news newsletter, issue no. 12, september 2004 · september 2004 • long-term care...

Long-Term Care NewsThe Newsletter of the Long-Term Care Insurance Section Published by the Society of Actuaries

September 2004, Issue No. 12

contents

Introduction

It's only in the last decade that most of the literature has been written on how thebrain functions and how we process communications. Because consumer reactionsto marketing communications will depend on the manner in which they’re

processed, an understanding of how the brain processes information can be veryuseful. Information processing refers to the process by which a stimulus is received,interpreted, stored in memory and later retrieved.

An appreciation of information-processing principles and findings can yield someimportant lessons for those interested in influencing consumer behavior. Althoughmarketing communications is perhaps the greatest beneficiary of what we know abouthow people process information, these lessons can be applied to many other areasincluding personal selling, package design, branding and training of salespeople.

What follows are recent findings of brain research having relevance to marketing andsales to boomer-plus customers. Author David B. Wolfe, a noted expert on develop-mental relationship marketing contributed much to the findings discussed.

Findings1) There are material differences between males and females in the architecture andfunctioning of their brains. This often leads to different responses to the same experi-ences. Females generally make greater use of right brain functions in thinkingprocesses, making them more subject to emotional arousal than males. However,research indicates that in later life, the gap between males and females in emotionalsensitivity narrows. Males become more intuitive and depend more than they didearlier in life on emotional reads of a situation to determine if it warrants further atten-tion.

MARKETING IMPLICATION: Logic in product messages generally works betterwith males than females. However, this doesn't mean qualitative differences in accu-racy of perceptions because females generally make more effective use of intuition, aright brain, emotionally based function. However, once a female experiences a favor-

Successful Marketing and Selling to Baby-Boomer-and-Older CustomersTwelve Recent Brain Research Discoveries Having Relevance to Successful CommunicationsBy James J. Gilmartin

continued on page 4

Successful Marketing and Selling to Baby-Boomer-and-Older Customersby James J. Gilmartin . . . . . . . . . . . . . . . . . .1

A Word From the Editorby Bruce Stahl . . . . . . . . . . . . . . . . . . . . . . . .2

Chairperson’s Cornerby James M. Glickman . . . . . . . . . . . . . . . .3

The Upside to Higher Persistency in LTCInsurance—Effects of Anti-SelectionAfter Policy Issueby John L. Timmerberg . . . . . . . . . . . . . . . . .7

Projecting Policy Persistency by Bruce Stahl . . . . . . . . . . . . . . . . . . . . . . .11

2 • Long-Term Care News • September 2004

Long-Term Care NewsIssue Number 12• September 2004

Published by the Long-Term Care InsuranceSection Council of the Society of Actuaries

475 N. Martingale Road, Suite 600Schaumburg, Ill 60173-2226

Phone: 847-706-3500 Fax: 847-706-3599

Web: www.soa.org

This newsletter is free to section members. To jointhe section, SOA members and non-memberscan locate a membership form on the LTCI Webpage at www.soaltci.org. Back issues of sectionnewsletters have been placed in the SOAlibrary and on the SOA Web site (www.soa.org).

2004 SECTION LEADERSHIPJames M. Glickman, ChairpersonSteve P. Sperka, Vice-ChairpersonVincent L. Bodnar, SecretaryRobert K. Yee, TreasurerPhilip J. Barackman, Council MemberBartley L. Munson, Council MemberMark D. Newton, Council MemberJames M. Robinson, Council MemberP. J. Eric Stallard, Council Member

Bruce A. Stahl, Newsletter EditorPenn Treaty Network America146 Lakeview Drive • Suite 203Gibbsboro, N.J. 08026PHONE: 856-566-1002FAX: 856-566-5165E-MAIL: [email protected]

Brad S. Linder, Co-editorGenRe | Life Health695 East Main Street • P.O. Box 300Stamford, Conn. 06904-0300Phone: (203) 352-3129Fax: (203) 328-5886E-mail: [email protected]

Mary Pienkowski, Graphic DesignerPHONE: 847-706-3548FAX: 847-706-3599E-MAIL: [email protected]

Clay Baznik, Director of PublicationsE-MAIL: [email protected]

Lois Chinnock, Sections ManagerE-MAIL: [email protected]

Facts and opinions contained herein are the soleresponsibility of the persons expressing them andshould not be attributed to the Society ofActuaries, its committees, the Long-Term CareInsurance Section or the employers of theauthors. We will promptly correct errors broughtto our attention.

Copyright © 2004 Society of Actuaries.All rights reserved.Printed in the United States of America.

Decisions, decisions, decisions. As a good chess playerconsiders the opponent’s movements in advance, so wewill do well to consider our clients’ thoughts and

emotions in advance. No matter what role we play in the long-term care insurance arena, we want to be able to recognize andhelp direct our clients’ thoughts and emotions toward favorabledecisions.

In our present rate stabilization environment where premiumrates have increased to levels that seem less affordable to many inthe individual market, Jim Gilmartin offers some interestingconsiderations in his article on successful marketing. Connectingmarketing techniques to mental processing activities, he identifies how the client isprocessing information so that the salesman can walk through the components ofeach sales presentation culminating with the client deciding to buy.

The policyholder continues to make decisions regarding his long-term care insurancepolicy after his initial purchase. In his article on adverse selection from voluntarylapses, John Timmerberg identifies a few situations in which the policyholder is onceagain confronted with a decision to protect his family from the consequences of hisrequiring long-term care. The agent knows that these policyholder decisions impacthis commissions, yet they also significantly alter the relationship between expectedpremium income and claim expenses.

These subsequent-to-purchase policyholder decisions are not all equal. The nature ofthe original sale, the policyholder’s previous decisions, his environment and hishealth status all influence the direction of the subsequent decisions. I discuss these inmy article on persistency, in order to demonstrate that there is value in recognizingthese decision differences. The value is in successfully planning for the future. ¯

A Word From the Editorby Bruce Stahl

EditorBruce Stahl

September 2004 • Long-Term Care News • 3

Welcome to this LTCI section newsletter,my last as your Section Chairperson for2004.

During the past five years, the LTCI Section hasprogressed from just a good idea, to one of themost active sections in the SOA. In fact, with ourattempts this year to expand membership beyondactuaries, we have taken the first step towardmaking this section the educational and informa-tional location of choice for the LTCI industry.Our annual intercompany conference (the fifthedition of which is being held at the Rosen Centrehotel in Orlando, January 23-26, 2005) has morethan doubled in size from just over 300 attendeesto more than 700 attendees in its first four years.This conference is now recognized by the LTCIindustry as the conference to attend, if you canonly attend one.

Last October, I set three primary goals for thesection to accomplish during my year as chairper-son:

• Increase LTCI Section membership morethan any other section with an emphasis onincreasing the number and participationlevel of the non-actuarial members.

• Make the LTCI Section Web site an indis-pensable tool for everyone in the LTCIindustry with more “hits” than any othersection Web site.

• Start the process of making the LTCISection the resource of choice for political,educational and research questions regard-ing LTCI and the LTCI industry.

Goal number one is a work in progress. Duringthe past 12 months, our section membership hasincreased by 15.5 percent, which was the largestgrowth rate in section memberships among the 17

sections. In addition, our non-actuarial member-ship is now over 25 percent and growing. By thistime next year, it should exceed 40 percent.

Goal number two has moved a little moreslowly, but has shown much progress. We havea totally redesigned Web site with informationorganized along eight different tracks (Actuarial,Claims, Compliance, Group, Management,Marketing, Operations and Underwriting). Inaddition, we now have our own Web site:www.SOALTCI.org that goes directly to the LTCISection Web pages, bypassing the rest of the SOAsite. However, there is still much work to beaccomplished to make this site an indispensabletool for the LTCI industry. I strongly encourageanyone with interest to contact me or anyone elseon the council to volunteer in this effort.

Goal number three has made the least progressof the three goals, although some significantprogress has been made there as well. Duringthe past year, both the IAA (InternationalActuarial Association) and the ManagingRetirement Assets Symposium have requestedand received assistance with projects involvingLTCI. Over time, as more organizations becomeaware of the LTCI Section, more will be accom-plished toward meeting this goal.

In order to continue moving forward on each ofthese goals, as well as to establish and implementadditional ones, we will need your help. Pleasecontact the section if you are willing to participatein these activities.

Finally, I would like to thank the many peoplewhose volunteer help made this year a successfulone for our section. With all the changes occur-ring in the LTCI industry, the work of this sectionin the next few years will become even moreimportant. ¯

Chairperson’s Cornerby James M. Glickman

James M. Glickman,

FSA, is president of

LifeCare Assurance

Company in Woodland

Hills, Calif. He can

be reached at

Jim.Glickman@

LifeCareAssurance.

com.

able insight, she may become as rational infurther processing of a matter as a male. It's justthat her right brain is a more formidable gate-keeper to the left brain than male brains generallyare.

2) Motivations do not originate in the consciousmind. The conscious mind is the executive offi-cer that, like a corporate CEO, makes decisions onneeds that have been framed at lower levels.Neurologist Richard Restak states in The BrainHas a Mind of Its Own, "We have reason to doubtthat full awareness of our motives may be possi-ble." Adds brain researcher Bernard Baars in Inthe Theater of the Brain, "Our inability to reportintentions and expectations simply reflect the factthat they are not qualitatively conscious."

MARKETING IMPLICATION: Answersconsumers give researchers about their motiva-tions are often incomplete or off the mark simplybecause people can only speculate about theirmotivations at deepest levels of the psyche.Creators of product messages need to becomemore intimately familiar than is typical with the"hidden drivers" of consumers' behavior thanconsumers, about which they have little explicitknowledge. These drivers tend to be stage-of-lifespecific. For example, young people generallyhave stronger outer-directed motivations relatingto social status than older people. Older people'smotivations tend to be qualitatively more experi-ential and less materialistic than youngerpeople's motivations.

3) People use different brain sites and mentalprocesses in answering researchers' hypotheti-cal questions than they use in real-lifesituations. Research respondents tend to drawmore heavily on the objective sequential reason-ing of the left brain than on the subjectiveemotional right brain in answering researchers'questions. This bias is reversed in favor of theright brain in reacting to product messages andmaking buying decisions.

MARKETING IMPLICATION: Research resultscan be improved by techniques that are moreeffective in divining consumers' implicit testi-monies that have not been distorted by undueinfluence from left-brain processing. The recenttrend toward studying consumers in their naturalliving and shopping environments is justified by

the finding that people process hypotheticalinformation differently than they do real-lifeinformation. Researchers need to make more useof indirect techniques to get behind the curtainsof consciousness.

4) Brain development is lifelong. How peoplementally process information changes from onedecade of life to the next. This alters how peopleview and connect with the external world (world-view). Language style preferences also changeover time. For example, youth and young adultsgenerally have a more assertive language stylethan older people.

MARKETING IMPLICATION: Productmessages will be more effective when expressedin the stage-of-life language style of the coremarket to which a message is primarilyaddressed.

5) Adolescent brains are significantly inferior toadult brains in reading facial expressions. Theolder people are, the more skilled they generallyare at reading facial expressions.


Successful Marketing and Selling to Baby-Boomer-and-Older Customers • from page 1

Creators ofproductmessages needto becomemore intimatelyfamiliar than istypical with the"hidden drivers"of consumers’behavior thanconsumers,about whichthey have little explicitknowledge.

MARKETING IMPLICATION: Productmessages depicting people should reflect aware-ness of the core audience's ability to read facialexpressions. For instance, older people's greatersensitivity to facial expressions means that facialexpressions should bear authentic connection tothe product and product message in oldermarkets. Younger consumers willgenerally be more concerned withwhat people are doing than with whattheir faces are saying.

6) As midlife (40+) approaches,people increasingly draw on right-brain functions. They begin relyingless on left-brain sequential reasoningand more on emotions—aka "gut feel-ings" or intuition.

MARKETING IMPLICATION:Product messages for people over 35should have more affect (emotionaltoning) than product messages foryounger people. Under 35, peopletend to have a stronger reasoning bias, thus prod-uct messages generally should implicitly orexplicitly promote concrete reasons for purchase.

7) Information entering the brain's cortex (outerlayers) is first processed mostly in the rightbrain. The right brain processes information assensory images rather than as words andnumbers. The left brain works in numbers andwords.

MARKETING IMPLICATION: To arouse thestrongest attention, product messages should berich in sensory stimuli. Even though the rightbrain can't process words, words can createsensory images, as every storyteller knows. Theolder a market, the more important it is to presenta product in story form.

8) Emotion, not reason, is the final arbiter indecision-making. Initial responses to informa-tion entering the brain are visceral. Changes inbody states (e.g., pulse, hormonal flow, salivaflow, body temperature, etc.) generate emotions.When a matter fails to generate emotions, aperson will not take action on it. (Brain patientswho have lost their emotional abilities whileretaining full powers of comprehension and

reasoning cannot make advantageous decisionsin which they have a personal stake in theoutcome.)

MARKETING IMPLICATION: A cardinal rulefor developing effective product messages is gowith the grain of the brain or "Lead with the

right; follow with the left." The only way to getinto a person's conscious mind is via the rightbrain. Again, sensory images are keys to gettinginto the right brain.

9) Gender tends to predispose responses tovoice-overs in broadcast advertising. For exam-ple, male voices are seen as more knowledgeablewhen describing technical attributes of a product,while female voices are seen as more knowledge-able when describing a product with references tolove, relationships and caring.

MARKETING IMPLICATION: Choose the voiceto match the content and delivery style of a prod-uct message.

10) Pictures of people in motion arouse thebrain more quickly than posed pictures.

MARKETING IMPLICATION: Avoid posedpictures like the plague. Motion conveys vitality.Posed pictures convey lifelessness. Posedpictures should be all but totally avoided inmarketing to older adults, although marketerscommonly use posed pictures for that market.


Successful Marketing and Selling to Baby-Boomer-and-Older Customers

continued on page 6

11) Each experience we have prompts the brainto create clusters of neurons (brain cells) withpredisposed responses to new but similar expe-riences. As the population of these dispositionalclusters increases, a person becomes more habitu-ated and reflexive in his or her responses. Thisdecreases sensitivity to external influences, likeadvertising, making a person more autonomous.

MARKETING IMPLICATIONS: Dispositionalclusters are the marketer's equivalent of "hotbuttons." The older we are the more hot buttonswe have. This is good news and bad news formarketers. First the bad news: It's harder tochange people's patterns after the early adultyears. Now, the good news: When a marketerhits a consumer's hot buttons, the deal is all butmade. The challenge is learning what those hotbuttons are. Fortunately, there is remarkableconsistency in the general nature of hot buttonsamong people in the same season of life.Knowledge of the developmental attributes ofconsumers in a given season will guide amarketer in making contact with their hotbuttons.

12) Initial determination of information rele-vance occurs unconsciously. When a person seesan ad, the right brain initially determines if the adhas personal relevance. The sequential reasoningprocesses of the left brain only go to work on thead after it has reached consciousness. The rightbrain conducts a process called information triageto reduce information flow to levels the consciousmind, with limited working memory (RAM) canhandle. The primary criterion is relevance to aperson's interests.

MARKETING IMPLICATION: Imagine havinga conversation in your office or at a social gather-ing when you hear your name come up inanother conversation not far from you. Yourbrain was hearing the other conversation allalong, but only when your name was mentioneddid it see fit to alert your conscious mind to theother conversation. That's what informationtriage is about. Creating product messages thatsurvive information triage is the biggest chal-lenge in marketing. It has become fashionable to

complain about advertising clutter. However, theclutter problem is in the brain, not on a televisionscreen or in a magazine. When a message hasrelevance to a person's interest, the right brainwill take note. When we talk about having a"double take,” we acknowledge the right brain'sability to pick up in a nanosecond something thathas relevance to our interests.

Don’t Muddle the MessageMarketing and sales professionals often pay littleattention to how the consumer thinks andprocesses information. Research has shown thatthe right hemisphere of the brain processesemotional information and the left processes logi-cal information such as product demonstrations.This knowledge can help to avoid blunders thatmight turn interest into disinterest.

Communications that evoke emotional responsestypically produce a high level of processing activ-ity in the right hemisphere. Unfortunately, manycommunications that draw the viewer, listener orreader into an emotional scene abruptly orquickly cut to product information. Deep insidethe brain, this action causes trouble. The righthemisphere is still highly active, making it diffi-cult for the brain to process words (the brain onlyprocesses images). In short, the timing canmuddle the message. ¯

Jim Gilmartin is president of Lombard, Ill.-basedComing of Age, Incorporated, and is a frequent speakerat sales and marketing conferences, leadership/management retreats and association meetings. Hecan be reached at 630-627-9600, or e-mail him [email protected]. You can also visitwww.comingofage.com for more information ontheir services.

Copyright ©Coming of Age, Incorporated


Successful Marketing and Selling to Baby-Boomer-and-Older Customers • from page 5

Jim Gilmartin is president

of Lombard, Ill.-based

Coming of Age,

Incorporated, an

integrated marketing

communications firm

specializes in helping

clients to increase

market share and profit

in boomer+ customer

markets. He can be

reached at 630-627-

9600, or e-mail him

at jimgilmartin@

comingofage.com


Alower level of voluntary lapses—orhigher persistency—is a topic of interestor concern for long term care insurance

(LTCI) profitability and pricing. The pricing ofLTCI is lapse-supported, meaning that voluntarylapse rates below expectations can lead to profitconcerns. But the flip side to high persistency isfewer opportunities for policyholder anti-selec-tion against the company. Higher persistencyshould result in improvements in experience asmeasured by policy year loss ratios and attainedage claim cost levels.

In this article, I discuss policyholder anti-selec-tion (after issue), develop a fairly simple model todemonstrate the impact on attained age claimcosts, and then compare results for three scenar-ios. Then I look at the variation in anti-selectionbetween higher and lower levels of lapses andone scenario includes the impact of additionalevent-specific lapses. I demonstrate that whenpricing products today in a lower lapse rate envi-ronment, actuaries may consider using favorableadjustments to their claim costs if using historicalexperience available from a higher lapse rateenvironment, all other factors held constant.

I credit William F. Bluhm’s article “CumulativeAntiselection Theory,” which is the foundation orinspiration for many of the ideas presented here.

Anti-selection in Action When using the term anti-selection, I am referringto the ability and tendency of LTCI policyholdersto make decisions regarding whether or not theywill voluntarily lapse their policy that reflect theirsuperior knowledge as to their health status andpotential future claim utilization. Generally, poli-cyholders who lapse may have made the personalprediction that their benefit utilization will below. Therefore, lapses tend to remove the healthi-est policyholders from the pool. This impacts theremaining pool, resulting in higher policy yearloss ratios and higher attained age claim costs, ascompared to the performance that would exist inthe absence of these lapses.

For example, I suggest that a policyholder whopurchased their policy seven or eight years ago,

and has since experienced two falls or a diagnosisof emerging Alzheimer’s disease and lives aloneis much less likely to lapse their policy ascompared to a policyholder who has few healthconcerns, remains active and lives with theirspouse. In addition to the policyholder’s knowl-edge regarding their current health status orpotential future ADL (activity of daily living)deficiencies, benefit utilization could be impactedby the presence of a spouse, the presence offamily nearby or other informal support services,relative accessibility of formal services, regionalvariations, the willingness of the policyholder toaccept the services or benefits available under

their policy and other factors. It is possible thatthe presence of a spouse or other factors competewith health status as a primary indicator of futureutilization. Although there are many factors thataffect future benefit utilization, I will use the term“health status” to indicate the combination of allpolicyholder-specific factors that affect benefitutilization.

The Upside to Higher Persistency in LTCInsurance—Effects of Anti-Selection After Policy IssueBy John L. Timmerberg

continued on page 8

The potential for policyholder anti-selectionexists at each point during the life of a LTCIpolicy when a premium payment is due. Thepolicyholder must decide whether or not theywill pay the premium thus keeping the policyactive, or not pay the premium and lapse thepolicy. And, if properly motivated, the policy-holder may cancel the policy between premiumpayments. Special circumstances or events whereone might expect additional event-specific lapsesand policyholder anti-selection are as follows:

• a rate increase notification is received; • the policyholder receives potentially unset-

tling news such as an insurance companyratings downgrade, sale, merger or exitfrom the LTC line of business;

• an increase in agent activity, with offers toreplace policies with newer policies thatmay have higher benefit value perpremium dollar;

• a communication regarding policyholderoptions as the result of a class action settle-ment, a regulatory intervention or someother unusual event that effects the policy-holder’s status; or

• a return of premium benefit (cash payable)matures and becomes payable.

In fact, additional policy lapses can occur at anypoint that the policyholder receives a communi-cation reminding them that their policy exists andis still active.

The ModelTo model the impact of anti-selection, I willdiscuss it in terms of the impact on the expectedclaim cost for one year. By claim cost, I am refer-ring to the attained age claim costs which aremeasured as claims incurred per dollar of dailybenefit amount exposed during the year, for aselected attained age. Attained age claims costcurves are used in both pricing and reserving(active life reserves) and increase rapidly withattained age, although with varying slopesdepending on company experience and theirinterpretation of the data.

I start with the following equation for the year inwhich the lapse occurs:

CCk = (1.00-VLRt)(HSA)(CCk) +(VLRt)(HSL)(CCk),

with variables defined as follows:

VLRt = voluntary lapse rate for policy duration t.CCk = claim cost for attained age k.HSA = average health status of continuing activepolicy for this year. HSL = average health status of lapsed policy forthis year.

As motivation for the equation, please note thatthe potential benefit utilization is split into twoparts: that for continuing policies and that whichwould have been contributed by policies thatlapsed. By assuming some level of health statusfor the lapsed policies (HSL) and using the appro-priate voluntary lapse rate for duration t (VLRt),we are able to calculate the health status forcontinuing active policies (HSA) that brings theclaim utilization total back into balance.

Assume HSL = 0.50 and VLRt = 0.04 and solve forHSA, then ...

1.00 = (1.00-0.04)( HSA) + (0.04)(0.50)

and HSA = 1.02083, or 2.1 percent higher as aresult of the lapses.

When assuming that HSL= 0.50 in the calculation,we are indicating that the health status of thelapsed policy in the year that they lapse is one-half the overall expected rate of lapsed andcontinuing policies combined. We will call this amoderate level of health status differential. Asanother example, if HSL= 0.25, that would indi-cate that the health status or expected benefitutilization rate of the lapsed policy is one fourththe overall expected rate of lapsed and continu-ing policies combined. We will call this a highlevel of health status differential.

The 2.1 percent increase calculated above is theimpact of the anti-selection in the year of thelapse. In addition, we assume that the effect ofthis year’s lapses continues to future years. Inour examples, we assume that the effect contin-ues but decreases 10 percent each year, runningoff within 10 years. As noted above, the differ-ence in future claims is a reflection of a number offactors, some of which are quite stable andpersistent. Thus, allowing the effect to run offover 10 years is reasonable. For each year, thecumulative impact of the anti-selection is the anti-selection for that year plus the multiplicativeimpact of prior years that are still running off.For example, in policy year three the effectincludes the first year effect for policy year three,the second year effect from policy year two andthe third year effect from policy year one.


The Upside to Higher Persistency in LTC Insurance • from page 7

The potential forpolicyholderanti-selectionexists at eachpoint during thelife of a long-term care insurance policywhen a premiumpayment is due.

The ResultsExhibit I shows three voluntary lapse rate scenar-ios. “Higher Lapse Rates, No Events” is from ahigher lapse rate era, possibly representative ofLTCI policies issued in the early 1990s. Columntwo, “Higher Lapse Rates, Two Events“ is fromthe same higher lapse rate era, but also includestwo years where event-specific effects on thelapse rate increased it by 400 basis points eachyear (a premium rate increase in year five and aninsurance company ratings downgrade in yearseven, for example). Column three, “LowerLapses, No Events” shows the expected lapse ratefor policies being issued during a low lapse rateera, which could be reflective of today, and noevents. Please see the “Long-Term CareInsurance Persistency Experience” joint study bythe SOA and LIMRA for information on thedifferences in lapse rates over recent time peri-ods.

Exhibit II shows the anti-selection impact of thelapse rate scenarios with some variation in theanti-selection intensity as measured by the differ-ential in health status. The first column showsthe cumulative anti-selection impact for the“Higher Lapses, No Events” scenario along withmoderate health status differential (HSL = 0.50).Column two shows the cumulative anti-selectiveimpact of the “Higher Lapses, Two Events” lapserates, along with high health status differential(HSL = 0.25). Column three shows the cumula-tive anti-selective impact of the “Lower Lapses,No Events” scenario along with moderate healthstatus differential (HSL = 0.50). The values

shown in the table are the percentage increase inthe claim costs resulting from the anti-selectionbrought about by both the underlying level ofvoluntary lapses and the additional anti-selectionproduced by the lapses resulting from events.

Column two shows the highest results, with anti-selection peaking at duration seven at 35.8percent and leveling off at 21.2 percent for alldurations, 17 and above. The anti-selectionimpact peaks at the year of the second event,which increased lapses by 400 basis points in thatyear. Column three shows the lowest results,with anti-selection peaking at durations four andfive at 6.5 percent and levels off at 4.3 percent atduration 13. Clearly, the impact of anti-selectioncan vary widely, depending on the level of lapses,additional lapses due to events and the healthstatus differential assumed.

As a potentially useful application, one canconsider the development of a claim cost curve asstarting from a theoretical curve that could exist ifthere were zero voluntary lapses. This “baseline”claim cost curve would be the lowest claim costcurve with all other curves resulting from non-zero lapses being at higher levels, all other factorsheld constant. To estimate the “baseline” claimcost curve, we take the values resulting from ouractual experience and then divide by the factorsresulting from our model. For example, if we hadstatistically credible claim cost values from resultsexperienced as described by the heading in


The Upside to Higher Persistency in LTC Insurance

Exhibit I

Policy Higher Lapse Rates Higher Lapse Rates Lower Lapse RatesDuration No Events Two Events No Events

1 15.50% 15.50% 5.50%

2 9.50% 9.50% 4.00%

3 7.00% 7.00% 3.00%

4 5.00% 5.00% 2.50%

5 4.50% 8.50% 1.50%

6 4.50% 4.50% 1.50%

7 4.50% 8.50% 1.50%

8 4.50% 4.50% 1.50%

9 4.50% 4.50% 1.50%

10 4.50% 4.50% 1.50%

11+ 4.50% 4.50% 1.50%

continued on page 10

...one canconsider thedevelopment of a claim costcurve as startingfrom a theoreti-cal curve thatcould exist ifthere were zerovoluntary lapses.

column two of Exhibit II, then we could divideour actual values by these factors to get back tothe “baseline” claim cost curve. Now, when pric-ing under new lapse rate assumptions, such asthose in column three of Exhibit II, we can takeour “baseline” claim costs and multiply by thecolumn three factors to create a set of expectedclaim costs resulting from this new lapse rateenvironment. This demonstrates that the newclaim costs to be used to price under the scenariodescribed by the column three headings are lowerthan those resulting from column two solely due to the difference in anti-selection, by theratios of column three divided by column two(1.029/1.138 for duration one, for example). Ofcourse, the claim costs would be adjusted forother product and underwriting differencesbetween the column two era product and thecolumn three era product.

Conclusion Although the impact of anti-selection may belower in today’s environment of lower lapse ratesand rate stability (implying that events that createadditional lapses will be rare), anti-selection byLTCI policyholders should be considered whendeveloping experience reports from historicaldata or selecting claim cost assumptions for pric-ing or reserving. Credible data from priorperiods should be adjusted to reflect differencesin lapse rates, including additional lapses result-ing from events. The selection of the health statusdifferential variable is an important considerationwhen modeling these effects, and actuariesshould consider how they might best develop acredible estimate for this variable. ¯


The Upside to Higher Persistency in LTC Insurance • from page 9

Exhibit II

Higher Lapse Rates Higher Lapse Rates Lower Lapse RatesPolicy No Events Two Events No Events

Duration Moderate HS Differential Higher HS Differential Moderate HS Differential

1 9.2% 13.8% 2.9%

2 13.9% 21.2% 4.8%

3 16.6% 25.6% 5.9%

4 17.7% 27.3% 6.5%

5 18.1% 32.2% 6.5%

6 18.2% 32.1% 6.4%

7 18.0% 35.8% 6.2%

8 17.5% 34.1% 6.0%

9 16.8% 32.1% 5.6%

10 15.8% 29.5% 5.2%

11 14.5% 26.5% 4.7%

12 13.9% 24.8% 4.5%

13 13.7% 23.7% 4.3%

14 13.7% 22.8% 4.3%

15 13.7% 22.0% 4.3%

16 13.7% 21.6% 4.3%

17 13.7% 21.2% 4.3%

18 13.7% 21.2% 4.3%

19 13.7% 21.2% 4.3%

20 + 13.7% 21.2% 4.3%

John L. Timmerberg,

ASA, MAAA is president

of Timmerberg &

Associates, Inc.,

providing actuarial

consulting specializing

in long-term care

insurance.


Many LTCI actuaries were watching theLTCI reserve assumption develop-ments in NAIC committees over the

past couple of years. One particularly sensitiveassumption was the policy termination rate. AsLTCI tended to be termination supported, someof the regulatory actuaries were naturallyconcerned that LTCI termination rates were muchsmaller than had been anticipated even four orfive years ago. They wanted to indiscriminatelyimpose very small termination rates on all insur-ers. Yet some company actuaries argued that theterminations were not always small, and variedfrom company to company.

We can understand why regulatory actuariesmight want to apply one standard to all compa-nies. As a pricing actuary for some small LTCIcarriers during the late 1990s, I was personallyconcerned about using the actual historiccompany experience for an issue-age priced prod-uct. I speculated that the smaller companyterminations may have been higher than those oflarger companies because of the company finan-cial ratings, and I considered that the smallcompanies might sell their LTCI to a highly ratedcompany.

Yet recent persistency studies reveal an explana-tion for the varying termination rates amongLTCI carriers, that is intrinsic to the businessitself, rather than to the company. Persistencyvaries significantly according to policy benefits,marital status, risk classification and distributionchannel. These factors are apparent in the 2004report, “Long-Term Care Insurance PersistencyExperience,” sponsored jointly by LIMRAInternational and The Society of Actuaries LTCExperience Committee. They are also apparent ina study that the Penn Treaty actuarial departmentrecently prepared. This study included regres-sions, separately by policy duration, of a varietyof factors on persistency.

Policy benefits. Reviewing the LIMRA/SOAstudy, most significant among the mainstreambenefit options appears to be the automaticincreasing benefit option, which tends to havehigher persistency than policies without it. Yethaving unlimited benefits or a lifetime benefit

period also experiences better persistency. ThePenn Treaty study had mixed results on both ofthese options, as the predictor coefficients did notbehave consistently by duration. For example,the increasing benefits option appeared to reducepersistency in early durations, but increase it inlater durations. One interpretation of this is thatthe high premium has more of an impact on theearly-duration persistency than it does on thelater.

Marital status. Both the LIMRA/SOA study andthe Penn Treaty study identify better persistencyfor married individuals. With a termination-supported product, this suggests there should bepressure for increases in premium rates formarried couples, possibly in the form of smallerspouse discounts. Spouse discounts are normallybased on lower morbidity for married couplesthan for singles, yet varying persistency assump-tions by marital assumptions may be appropriateas well.

Risk classification. The small amount of data inthe LIMRA/SOA study identifies a lower lapserate for substandard policies than for standard.The Penn Treaty experience includes much moreexposure in substandard classes, and it suggeststhat total terminations are significantly higher forsubstandard classes. Penn Treaty data does notidentify mortality and voluntary lapses sepa-rately. The substandard classes may have muchhigher mortality. Yet a reasonable conclusion isthat substandard classes experience lower persis-tency.

Distribution channel. The LIMRA/SOA studyprovided confirmation of what the industry hasknown for some time. The LTCI business that issold through captive agents experiences higherpersistency than that sold through independentproducers.

Because persistency varies significantly by thecharacter of the business, actuaries will do well toconsider the intrinsic character variations in theirprojections, whether they be for pricing, financialmodeling, asset adequacy analysis or GAAPrecoverability testing. ¯

Projecting Policy Persistencyby Bruce Stahl

Bruce Stahl is senior

vice president and

chief actuary at

Penn Treaty Network

America in Gibbsboro,

NJ. He can be

reached at bstahl@

penntreaty.com

475 N. Martingale Road, Suite 600Schaumburg, Ill 60173-2226

Articles Needed for the NewsYour help and participation are needed and welcomed. All articles will include a byline to giveyou full credit for your effort. Long-Term Care News is pleased to publish articles in a secondlanguage if a translation is provided by the author. If you would like to submit an article, pleasecall Bruce Stahl, editor, at 856-566-1002.

Long-Term Care News is published quarterly as follows:

Publication Date Submission DeadlineDecember 2004 Monday, September 20, 2004

Preferred FormatIn order to efficiently handle articles, please use the following format when submitting articles:

Please e-mail your articles as attachments in either MS Word (.doc) or Simple Text (.txt) files. Weare able to convert most PC-compatible software packages. Headlines are typed upper andlower case. Please use a 10-point Times New Roman font for the body text. Carriage returns areput in only at the end of paragraphs. The right-hand margin is not justified.

If you have questions, or if you must submit in another manner, please call Mary Pienkowski, 847-706-3548, at the Society of Actuaries.

Please send a copy of the article to the following:

Co-editor

Brad Linder

Editor

Bruce Stahl

Brad S. Linder, ASA, MAAA,FLMI, ACS, Co-editorGenRe | Life Health695 East Main Street • P.O. Box 300Stamford, Conn. 06904-0300Phone: 203-352-3129Fax: 203-328-5886E-mail: [email protected]

Bruce Stahl, Newsletter Editor Penn Treaty Network America146 Lakeview Drive • Suite 203Gibbsboro, N.J. 08026PHONE: 856-566-1002FAX: 856-566-5165E-MAIL: [email protected]

long-term care news newsletter, issue no. 12, september 2004 · september 2004 • long-term care...

Documents